OPTICAL REVIEW Vol. 21, No. 5 (2014) 621–627
Failure Diagnosis of Organic Photovoltaic Using Electro-Optic Probe Jun KATSUYAMA1 , Kazuki MATSUMOTO1 , Ryo SUGIYAMA1 , Shinya HASEGAWA1 , Mitsuru SHINAGAWA1 , and Yoshiki YANAGISAWA2 1 2
Faculty of Science and Engineering, Hosei University, Koganei, Tokyo 184-8584, Japan Innovation Headquarters, Yokogawa Electric Corporation, Musashino, Tokyo 180-8750, Japan
(Received January 8, 2014; Accepted May 15, 2014) This paper describes the measurement of an organic photovoltaic using an electro-optic probe. We verify the validity of the electro-optic probe by comparing it with a conventional electric probe. The electric field distribution of the organic photovoltaic is examined on the basis of results of the test board measurement using the electro-optic probe and results of the test board simulation using an electromagnetic field simulator. We succeed in failure diagnosis of the organic photovoltaic in three failure modes. We found that the organic photovoltaic has a failure cell, and the electro-optic probe is successfully applied to failure diagnosis of organic photovoltaic. # 2014 The Japan Society of Applied Physics Keywords: electro-optic probe, organic photovoltaic, failure diagnosis, electric field distribution, electromagnetic field simulation
1.
Introduction
Floating and noncontact measurement
An organic photovoltaic (OPV) is a solar battery which is applied to a wide range of areas such as car roofs, curved surfaces of buildings, and clothes because it is thin, light, and flexible.1–3) It has been reported that the OPV can be used not only outdoors but also indoors.4) Therefore, it is a useful device for energy harvesting. The OPV consists of several solar cells connected in series. These cells are made from bulk heterojunction materials comprising p-type and n-type organic semiconductors.5,6) Investigations have shown that improvement of conversion efficiency is affected by organic semiconductor solution blend rate and the temperature of the production process.7,8) The OPV can be mass-produced by printing using a roll-to-roll process. The roll-to-roll process for the OPV is simpler and more cost-effective than the wafer dicing process for silicon solar battery.9) In-line inspection after every process is better than checking after all of the processes because the inspection system must immediately feed information back to the deposition system and stop the system when inferior quality or failure was detected. Therefore, precise measurement of the OPV is needed in roll-to-roll process. However, a probe cannot come in contact with the signal line and ground line during the process because OPV substrate is always moving, as shown in Fig. 1. For this reason, the OPV should be estimated by a floating and noncontact measurement method. An electro-optic (EO) probe is a noncontact10,11) and noninvasive12) probe. As application of the EO probe, GHzband high-speed LSI measurement10) and intra-body communication13–15) have been reported. The EO probe does not need reference potential because it directly measures electric field.16,17) It means that the EO probe need not come in contact with the ground. The EO probe can be used for noncontact low-disturbance measurement of electric field. Against this background, to solve the measurement problem encountered in the OPV roll-to-roll process, we use the EO probe.
Deposition system
Feedback OPV substrate
Excitation light
Fig. 1. In-line inspection of OPV in roll-to-roll production process.
First, we explain the configuration and performance of the EO probe. Next, we explain the measurement and simulation results of electric field distribution above the OPV using the EO probe. Finally, we discuss the experiments and electromagnetic field simulations of failure diagnosis. 2.
Electro-Optic Probe
2.1 Principle of EO probe This section explains the configuration and performance of the EO probe. The EO probe can be used to measure electric field with an EO crystal and laser light, as shown in Fig. 2. There are electric fields from the signal electrode to the ground electrode passing through the EO crystal. The laser beam from a laser diode (LD) passes through the EO crystal and is reflected by a high-reflectivity coating evaporated on the bottom surface of the EO crystal.10) The laser beam does not enter into the device under test; thus, it does not affect the device under test. The laser beam polarization changes when the electric field passes through the EO crystal depending on the strength of the electric field.18) The polarization change of the laser beam is converted to an intensity change as it passes through a polarizer. The intensity change is detected by a photodiode 621
J. KATSUYAMA et al.
Intensity change Polarizer
Polarization change EO crystal
Electric field
Ground electrode
Fig. 2.
Signal electrode
Faraday rotator Polarizing beam splitters Wave plates
Control VOUT unit
EO probe VIN
Oscilloscope 1k
Ground electrode
Measurement schematic of electro-optic probe.
1M
10k 100k Frequency [Hz]
Fig. 4. Frequency characteristics of EO probe and schematic diagram of experimental setup.
1
Control unit
Laser diode Photodiode Photodiode EO crystal Mirrors
VOUT / V IN [5dB / div.]
OPTICAL REVIEW Vol. 21, No. 5 (2014)
Optical fibers
EO probe Voltage [V / V MAX]
622
0.75 h 0.5
ΔV
0.25
Δh
Probe head 0
Fig. 3. Configuration of EO probe.
(PD). The PD generates photocurrent and converts optical signals into electric signals. The EO probe does not need reference potential because it directly measures electric field. Therefore, the EO probe need not come in contact with the ground line and it can be applied to noncontact measurements. 2.2 Configuration and performance of EO probe Figure 3 shows the configuration of the EO probe.19) The EO probe has a probe head and a control unit, which are connected by optical fibers. Active devices such as LD, PD, and semiconductor devices are far from the probe head. They are located in the control unit. There are passive devices such as the EO crystal, wave plates, and the polarizer in the probe head close to the device under test, and there is no metal. Therefore, the EO probe can precisely measure electric field with an extremely low disturbance. A cadmium telluride (CdTe) crystal is used as the EO crystal.12) The direction of the electric field detected depends on the cut surface of the EO crystal. The CdTe crystal with the (100) cut can detect a longitudinal electric field. First, the frequency characteristics of the EO probe are measured using a 50 microstrip line with a signal line width of 2 mm. The frequency characteristics of the EO probe and the schematic diagram of experimental setup are shown in Fig. 4. The frequency range from 10 to 40 kHz show flatness within 0.5 dB. The effective 3-dB-decrease high-pass cutoff frequency of the EO probe is 5 kHz. These characteristics are dependent on the resistivity of CdTe used
0
0.5
1 h [mm]
1.5
2
Fig. 5. Dependence of detected signal strength on gap between probe tip and microstrip line.
as the EO crystal in the EO probe. The resistivity of CdTe is not so high, and carriers move in CdTe. When the electric fields pass through CdTe, carriers locate in a certain area and inverse electric fields are generated by the location of carriers. The movement of carriers follows the change in electric field when the frequency of a signal is decreased. As a result, the sensitivity of the EO probe deteriorates in lowfrequency regions. The effective 3-dB-decrease low-pass cutoff frequency of the EO probe limited by the receiving circuit is 1 GHz.18) Second, signal strength was measured by varying the vertical gap h between the probe tip and the microstrip line. The dependence of detected signal strength on the gap h is shown in Fig. 5. When the vertical movement of the probe tip is 0:5 h 1:5 mm, the variations of detected voltage, V , are less than 20%. Third, signal strength was measured by varying the distance d between the center of the probe tip and the center of the microstrip line. The vertical gap h between the probe tip and the signal line was 1 mm. The EO probe position was fixed, and the microstrip line on the X-stage was moved. The dependence of detected signal strength on the distance d is shown in Fig. 6. When the probe position is 2 mm from the signal line edge, the detected voltage decreases by 50%. We defined that the spatial resolution of the EO probe is 2 mm.
OPTICAL REVIEW Vol. 21, No. 5 (2014)
J. KATSUYAMA et al.
1
623
20 cells
Voltage [V / V MAX]
EO probe 0.75
d
50 % down
213 mm
13 mm
0.5
2 mm
0.25 0
Spatial resolution 0
2
d [mm]
298 mm 4
6
P-type semiconductor
Signal line edge
Fig. 6. Spatial distribution of signals detected by EO probe above microstrip line.
Positive electrode
N-type semiconductor Negative electrode
Fig. 8. Photograph and cross section of OPV. EO probe RBW=1 Hz
500 mV input
20 cells
1 mm gap
200 mm 40 dB
Lead
298 mm PD
(a) Fig. 7. Minimum detectable voltage using spectrum analyzer without average.
Finally, the signal waveforms of the EO probe detected using a spectrum analyzer when the vertical gap h between the probe tip and the signal line is 1 mm is shown in Fig. 7. The 20 kHz 500 mV sinusoid signal was measured using a spectrum analyzer whose resolution bandwidth is 1 Hz without average. It was found that the signal-to-noise ratio of the EO probe is 40 dB. Therefore, we can distinguish between signal and noise when the input signal is more than 5 mV. Therefore, the sensitivity of the EO probe is defined as 5 mV when the vertical gap h between the probe tip and the signal line is 1 mm and resolution bandwidth is 1 Hz. 3.
Experimental Methods
3.1 Experimental setup Figure 8 shows a photograph and a cross section of the OPV. The OPV has 20 cells of 13 213 mm2 size. The interval between cells is 2 mm. All cells are connected in series. These cells are made from bulk heterojunction materials comprising p-type and n-type organic semiconductors. When light is input into these cells, they generate electromotive force and work as batteries. When there is no light, they work as capacitors between the positive and negative electrodes. The electric field distribution above
PD
Lead wire
(b) Fig. 9.
(a) Photograph and (b) structure of test board.
the OPV was measured by the EO probe by floating and noncontact measurement. The open circuit voltage is 500 mV per cell. The EO probe can measure the electric field of the OPV because the probe system has 5-mV sensitivity, as shown in Fig. 8. For the experiment, we used an electromagnetic field simulator, Agilent EEsof EDA EMPro 2013. 07, with Advanced Visualization. The structure of the OPV is complicated and it is difficult to create a simulation model of the OPV. Therefore, a test board was made, and it is shown in Fig. 9(a). It was a double printed board made of glass epoxy. There were 20 cells in a pattern of 13 200 mm2 . The interval between cells was 2 mm. All cells were connected by lead wires in series, as shown in Fig. 9(b). The PD was connected to the cell. When light is input to these PDs, they generate electromotive force. When there is no light, they work as capacitors between electrodes.
624
OPTICAL REVIEW Vol. 21, No. 5 (2014) Control unit
J. KATSUYAMA et al. Spectrum analyzer
OPV EO probe
EO probe
Z
Y
OPV
X LED driver
LEDs Moving
X-Stage
Rail
X-Stage
Rail
(a)
Z
(b)
Fig. 10.
(a) Schematic diagram and (b) photograph of experimental setup.
4th
6th
Y X
2nd
3rd
Fig. 11.
5th
7th
8th
9th
10th 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th
Simulation results of electric field distribution above test board using EMPro.
The electric field distribution above the test board was measured by the EO probe and simulated by the EMPro. A schematic diagram of the experimental setup is shown in Fig. 10(a), and a photograph of the experimental setup is shown in Fig. 10(b). A light-emitting diode (LED) was driven by an LED driver. The test light from the LED excited the OPV. The EO probe measured electric field distribution above the OPV. The EO probe was fixed on the X-stage, and it was moved on a rail. The measurement points were at the center of every cell. The EO probe was connected to a control unit, and the output signal of the control unit was measured by the spectrum analyzer. Considering the flatness of frequency range from 10 to 40 kHz, the test light should be modulated in this frequency span. The test board was simulated by an 11 kHz signal limited by EMPro with Advanced Visualization. Thus, the test light was modulated by an 11 kHz square wave signal. The vertical gap between the probe tip and the OPV is varied by about 0.5 mm because the upper surface of the OPV is not flat. Thus, the gap was set at 1 mm. 3.2
Electromagnetic field simulation and measurement of test board The electric field distribution above the test board was simulated by EMPro. Figure 11 shows the simulation results of the X–Z plane electric field distribution, which is Y ¼ 100 mm using EMPro. The origin of the Y-coordinate is defined as the front side edge of the test board. The
1 Voltage [V / VMAX]
1st
EO probe Simulation Coaxial cable probe
0.75 0.5 0.25 0 0
100 200 Probe position [mm]
300
Fig. 12. Experimental and simulation results of electric field distribution above test board.
potential of the rightmost cell is the highest because it is the sum of all electromotive forces generated by all cells connected in series. Therefore, there are electric fields from the 16th–20th cells to the 1st–5th cells, as shown in Fig. 11. The electric field distribution above the test board was measured by the EO probe with same experimental setup, as shown in Fig. 10. Figure 12 shows the experimental results and simulation results of the electric field distribution above the test board. The vertical axis shows the detected signal strength normalized by the maximum detected voltage. The horizontal axis shows the probe position from the left edge
OPTICAL REVIEW Vol. 21, No. 5 (2014)
X-component is dominant
0.75 0.5 7th cell 0.25 0
200 100 Probe position [mm]
0
Fig. 13. Direction of electric fields above test board.
3.3 Measurement of actual OPV device Figure 14 shows the experimental results of the electric field distribution above the OPV using the EO probe. It does not show a symmetric V-shaped distribution. At the point of the 7th cell, the measured voltage decreases sharply. We supposed that the OPV failed, and failure diagnosis was performed in the succeeding section. Failure Diagnosis
For a failure diagnosis, we planned three failure modes: 1) connecting a positive electrode to the next positive electrode corresponding to the bridge mode; 2) disconnecting a positive electrode to the next negative electrode corresponding to the open mode; 3) removing a PD from a cell corresponding to the PD defect mode. The electric field distribution above the test board in three failure modes was measured using the EO probe, and they are simulated using EMPro.
300
Fig. 14. Experimental results of electric field distribution above OPV using EO probe.
Bridge 6th
7th
8th
Fig. 15. Structure of test board in bridge mode.
Bridge point 1 Voltage [V / VMAX]
of the test board. Firstly, simulation results are discussed. There are electric fields, as shown in Fig. 13. At both ends of the test board, the Z-component of the electric field is dominant. At the center of the test board, the X-component of the electric field is dominant. The simulation value is the strength of the Z-component of the electric field. It means that the electric field distribution graph above the test board shows a symmetric V-shaped distribution. Next, measurement results are discussed. The EO crystal in the EO probe can detect a longitudinal electric field, thus, it can measure the Z-component of the electric field. Therefore, the measurement results ( ) agree with the simulation results ( ), and they show symmetric V-shaped distributions. We found that the EO probe can precisely measure the electric field distribution above the test board. We verified the validity of the EO probe. In addition, a coaxial cable probe was used for measurement of the electric field distribution above the test board. It is an electrical probe made from a coaxial cable and a copper tip of 2 2 0:3 mm3 . One end of the core wire was connected to the copper tip. It was capacitively coupled with the test board. The coaxial cable probe disturbed the electric field above the test board because the coaxial cable probe made of metal was close to the test board. For this reason, the distribution ( ) is not symmetric V-shaped. The coaxial cable cannot be applied to precise measurement.
4.
625
1
Z-component is dominant
Voltage [V / VMAX]
Z-component is dominant
J. KATSUYAMA et al.
0.75 0.5 0.25 0 0
Meas. Sim. 200 100 Probe position [mm]
300
Fig. 16. Experimental and simulation results of electric field distribution above test board in bridge mode.
First, the test board in the bridge mode was diagnosed. The 6th cell was connected to the 7th cell by a lead wire, as shown in Fig. 15. It means that the PD of the 7th cell is shorted. Figure 16 shows the experimental results of the electric field distribution above the test board in the bridge mode using the EO probe and the simulation results of the electric field distribution above the test board in the bridge mode using EMPro. The measurement results ( ) agree with the simulation results ( ). The electromotive force generated by the 7th cell decreased because the PD of the 7th cell was shorted. The potential of the 7th cell was low and there were more electric fields toward the 7th cell.
626
OPTICAL REVIEW Vol. 21, No. 5 (2014)
6th
7th
J. KATSUYAMA et al.
8th
6th
7th
PD defect
Open Fig. 17.
Structure of test board in open mode.
Fig. 19.
Structure of test board in PD defect mode.
PD defect point
Open point 1 Voltage [V / V MAX]
Voltage [V / V MAX]
1 0.75 0.5 0.25 0
8th
Meas. Sim. 0
100 200 Probe position [mm]
0.75 0.5 0.25 0
300
Meas. Sim. 0
200 100 Probe position [mm]
300
Fig. 18. Experimental and simulation results of electric field distribution above test board in open mode.
Fig. 20. Experimental and simulation results of electric field distribution above test board in PD defect mode.
Therefore, more electric fields were detected at the bridge point. Second, the test board in the open mode was diagnosed. The open mode means disconnection of series connection of cells. The lead wire connecting the 6th cell to the 7th cell was cut, and the cells were disconnected, as shown in Fig. 17. From the open point, the test board was disconnected into 6 cells (1st–6th cells) and 14 cells (7th–20th cells). Figure 18 shows the experimental results of the electric field distribution above the test board in the open mode using the EO probe and the simulation results of the electric field distribution above the test board in the open mode using EMPro. The measurement results ( ) agree with the simulation results ( ). Two V-shaped distributions are shown in Fig. 18 because the test board was disconnected. Finally, the test board in the PD defect mode was diagnosed. The PD defect mode means a fault of a cell. The PD of the 7th cell was removed, as shown in Fig. 19. The positive electrode and the negative electrode of the 7th cell are capacitively coupled. Figure 20 shows the experimental results of the electric field distribution above the test board in the PD defect mode using the EO probe and the simulation results of the electric field distribution above the test board in the PD defect mode using EMPro. The simulation value is the strength of the Z-direction electric field. The EO probe measured the Z-direction electric field because the EO probe is a longitudinal sensor. Therefore, we supposed that the measurement results ( ) agree with the simulation results ( ). However, they disagree to a certain extent.
On the basis of these failure diagnoses, a condition of the OPV is discussed. It was verified that the electric fields toward a failure cell are large and the electric fields toward a cell next to the failure cell are small. The voltage measured at the point of the 6th cell increases and the voltage measured at the point of the 7th cell decreases, as shown in Fig. 14. From the results, we found that the faulty cell is the 6th cell of the OPV, and the OPV in the bridge mode or PD defect mode is failed. For distinguishing failures in these modes, improvement of the EO probe performance and a more precise simulation model of the test board are needed. 5.
Conclusions
In conclusion, the purpose of this paper was to estimate of the OPV by in-line inspection in the roll-to-roll process with floating and noncontact measurement. We tried using the EO probe for electric field measurement of the OPV. We verified that the EO probe can be applied to measuring electric field distribution above the OPV. The electric field distribution of the OPV was measured using the EO probe and the electric field distribution of the test board was simulated using an electromagnetic field simulator, EMPro. Failure diagnosis in the three fault modes was performed. We found that the faulty cell is the 6th cell of the OPV, and the OPV in the bridge mode or PD defect mode is failed. The condition of the OPV can be supposed by measuring electric field distribution. We plan to improve the frequency characteristics of the EO probe by modifying the receiving circuit or changing the EO crystal and apply the EO probe to measurement in various situations.
OPTICAL REVIEW Vol. 21, No. 5 (2014) Acknowledgments The authors would like to thank Ryo Yoshikawa, Tatsuhiro Akiyama, and Ryo Saito for their help with experiment. They are grateful to Yoshinori Matsumoto, Masato Ishikawa, Akishige Ito, and Toyoaki Hamaguchi for their useful discussion. They also thank Katsuya Ikezawa and Takashi Tsubota for their continuous encouragement. References 1) C. W. Tang: Appl. Phys. Lett. 48 (1986) 183. 2) C. Lungenschmied, G. Dennler, H. Neugebauer, S. N. Sariciftci, M. Glatthaar, T. Meyer, and A. Meyer: Sol. Energy Mater. Sol. Cells 91 (2007) 379. 3) R. Tipnis, J. Bernkopf, S. J. Jia, J. Krieg, S. Li, M. Storch, and D. Laird: Sol. Energy Mater. Sol. Cells 93 (2009) 442. 4) M. Niggemann, B. Zimmermann, J. Haschke, M. Glatthaar, and A. Gombert: Thin Solid Films 516 (2008) 7181. 5) P. Peumans, S. Uchida, and S. R. Forrest: Nature 425 (2003) 158. 6) J. K. Lee, W. L. Ma, C. J. Brabec, J. Yuen, J. S. Moon, J. Y. Kim, K. Lee, G. C. Bazan, and A. J. Heeger: J. Am. Chem. Soc. 130 (2008) 3619. 7) W. Ma, C. Yang, X. Gong, K. Lee, and A. J. Heeger: Adv.
J. KATSUYAMA et al.
627
Funct. Mater. 15 (2005) 1617. 8) G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, and Y. Yang: Nat. Mater. 4 (2005) 864. 9) S. Logothetidis: Mater. Sci. Eng. B 152 (2008) 96. 10) M. Shinagawa and T. Nagatsuma: IEEE Trans. Instrum. Meas. 43 (1994) 843. 11) K. Tsuru, K. Murata, Y. Ohtani, Y. Nagai, and M. Shinagawa: SBMO/IEEE MTT-S IMOC’99 Proc., 1999, p. 648. 12) H. Togo, N. Kukutsu, N. Shimizu, and T. Nagatsuma: J. Lightwave Technol. 26 (2008) 2700. 13) M. Shinagawa, A. Sasaki, A. Furuya, H. Morimura, and K. Aihara: Jpn. J. Appl. Phys. 52 (2013) 09LA03. 14) A. Sasaki, M. Shinagawa, and K. Ochiai: IEEE Trans. Instrum. Meas. 58 (2009) 457. 15) Y. Song, K. Zhang, Q. Hao, and J. P. Rolland: Opt. Express 20 (2012) 13488. 16) J. A. Valdmanis, G. Mourou, and C. W. Gabel: Appl. Phys. Lett. 41 (1982) 211. 17) K. J. Weingarten, M. J. W. Rodwel, and D. M. Bloom: IEEE J. Quantum Electron. 24 (1988) 198. 18) T. Nagatsuma, M. Shinagawa, M. Yaita, and K. Takeya: IMTC ’94 Proc., 1994, p. 1476. 19) M. Shinagawa, T. Nagatsuma, K. Ohno, and Y. Jin: IEEE Trans. Instrum. Meas. 50 (2001) 1076.
